Expansion Intelligence Best Practices
Tips for getting the most from expansion recommendations — when to act, how much to order, and common mistakes to avoid.
Start with low-risk, high-confidence candidates#
Your first expansion should be a candidate with high confidence and low cannibalization risk. This lets you validate the feature before making bigger bets. Variant extensions of bestsellers are typically the safest starting point.
Order small, test fast#
Use the recommended opening order on each candidate card as your starting quantity. This is calibrated to give you enough stock to test demand without over-committing capital. If it sells through, reorder at a larger quantity.
Run analysis after major changes#
Re-run your expansion analysis after:
- Adding a new data source or uploading fresh sales data
- Seasonal transitions (pre-holiday, post-holiday)
- Launching a new product (the model recalculates cannibalization)
- Significant changes in your product range
Common mistakes to avoid#
- Ignoring cannibalization risk — launching a product that steals from your bestseller can reduce total margin
- Over-ordering trend-led products — trends fade fast, start small
- Skipping the test phase — even high-score candidates should be tested before a full commitment
- Only looking at margin — a 40% margin product that sells 5 units is worth less than a 20% margin product that sells 500
Frequently Asked Questions
Was this article helpful?
Still stuck? Email our support team.